=== twister_ is now known as twister [01:22] how can I force /etc/rc4.d/S99* to wait for S98 scripts to finish before firing off on boot? [01:25] would there be a better place to ask an upstart-related question? [01:26] nm, there is an #upstart channel.... === twister_ is now known as twister === unreal_ is now known as unreal [02:13] hey guys, anyone know a command to probe your local network? [02:13] thanks [02:14] you'll have to be more specific [02:14] for command line [02:14] i want to ssh into my box at work [02:14] then into another box [02:14] but i dont know its ip [02:14] nmap is the best port scanner I know of [02:15] it isn't included by default, you will have to install it [02:15] thanks qman__ ill try it [02:16] the command you'll want to use should look like [02:16] sudo nmap -sS -p 20 192.168.1.0/24 [02:16] err, -p 22 [02:16] and obviously use your network range [02:17] thank you qman_ [02:18] Error #487: Your port specifications are illegal. [02:19] that exact command works here [02:19] hmm [02:19] oh my bad [02:19] i forgot the -p 22 [02:19] i changed the /24 lol [02:20] heh, yeah, that's CIDR notation [02:21] thanks qman__ worked great [02:21] now i just have to figure out which ones mine lol [02:22] is there a way to resolve host names? [02:22] well, you could reverse dig each one [02:23] there might be an option for that in nmap, but there's a lot of them [02:23] prepare to read the manual for a while [02:24] could you briefly explain reverse dig please? [02:24] just did a fresh install and my nic isnt working, at all. i cant get it in ifconfig [02:25] dig -x 1.2.3.4 [02:25] thanks qman__ :0 [02:25] :)* [02:38] so i am a total newb at this, i have a basicly non configured apache server running on one pc and thats all i can handle. I am curious as how i would go about getting more then one server serving data, sort of like dual core only with servers. [02:39] dual core with servers? [02:40] i mean like so when i request a file from an IP, that multiple servers serve the data back to the requester [02:41] resno ^^ [02:42] nerdy_kid: im not sure what you are trying to do. so im hoping someone smarter will respond [02:42] are you saying multiple phyisical servers? [02:43] resno yeah serving the same IP [02:44] you mean like doing load balancing? [02:44] resno yesss [02:44] exactly :) [02:44] oh, i havent done anything with that. ive been meaning to try it out, but havent yet [02:45] is it a largly complex thing to do? [02:45] i havent done anything with it, so you are asking the wrong person. google or #apache2 would be a good place to start [02:46] resno ok thanks :) at least now i know what it is called :) [02:46] sure, thatll give you something to search for [02:46] resno yeah thanks :D === twister_ is now known as twister [03:00] ah, much better [03:00] anyone know how to output to file with wget? [03:02] what are you looking to output? [03:03] im wgetting a page, then comparing it to another file [03:03] i want wget to output to a file [03:03] i tried wget -o file [03:03] but no workie [03:04] just tried wget sitename >> file and (|) file with no success either [03:05] what about wget file && diff? [03:06] fried_pe1guin: i did try that [03:06] kvm question: do I want VT support enabled in the bios for kvm? [03:06] ah got it to work [03:06] silly me === fried_pe1guin is now known as fried_penguin [03:06] big O not little o [03:18] ok i found the problem, there is a & in the url [03:18] so when ever i try to run wget it tries to run everything after the & as another command [03:18] anyway to get it to stop that? [03:19] Put it in quotes. [03:19] The whole URL. [03:20] thanks ziesemer_ [03:20] Single will work best. [03:21] awesome thanks ziesemer_ [03:23] worked like a charm ;p [03:23] still dont see how charms work tho... [03:29] anyone use google reader? === gospch_ is now known as gospch [04:19] fried_penguin, yes [04:31] New bug: #580800 in postfix (main) "package postfix 2.7.0-1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 75" [Undecided,New] https://launchpad.net/bugs/580800 [04:31] need a hand with fstab === therian1 is now known as mellisagibson === mellisagibson is now known as therian_ [06:46] qman__ my SFTP apparently is working good [06:48] i need a little orientation this article http://is.gd/c9lXs have several times this "chroot()" what is that is a function??? [06:53] Is it possible to have primary and backup ubuntu repos? i.e. update from the local mirror .... if the local mirror fails to respond, try ubuntus mirror [08:11] New bug: #580819 in clamav (main) "package clamav-base 0.96 dfsg-2ubuntu1.2 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/580819 [08:13] the good of filezilla is that its shows you with messages what is exactlyy happening woth your connection [08:14] ouch!!! wrong chat [09:15] I installed the package [09:15] I installed the package 'postgresql', purged it, an reinstalled it [09:15] now the configuration files in /etc/postgresql are missing [09:15] how can I regenerate them? [09:16] or how can I make aptitude install them again? [09:28] should be there when you reinstalled [09:33] RoyK: the truth is much, much uglier >( [09:34] I had to purge the postgresql packages (and its dependencies), the needed to manually delete /etc/postgresql, /etc/postgresql-common, delete the user and the group 'postgresql' and run 'dpkg -P' on all the postgresql packages [09:34] then, and only then, a new installation of postgresql-8.4 would create the files in /etc/postgresql [09:47] what about postgresql-common or whatever it's named? [09:48] http://karlsbakk.net/top-16.png <-- nice [09:55] RoyK: /etc/postgresql-common or the package postgresql-common? anyway the directory /etc/postgresql-common was recreated as it should be without the "purging stunt". just /etc/postgresql was the problem [09:57] joschi: which distro version is this? [09:59] karmic [09:59] RoyK: https://bugs.launchpad.net/ubuntu/+source/postgresql-8.2/+bug/108296 suggests that this problem exists for a long time now... [09:59] Launchpad bug 108296 in postgresql-common "missing /etc/postgresql/ after uninstall/reinstall" [Undecided,Invalid] [10:01] joschi: I can upload the default /etc/postgresql* files somewhere, for lucid, but I don't have access to any Karmic boxes from here [10:01] RoyK: thank you but that's not necessary [10:01] k [10:02] RoyK: after purging everything related to postgresql (e. g. manually removing the user and the group "postgres"), aptitude recreated the directory [10:02] ok [10:02] RoyK: but I find it quite disturbing that this bug (?) survived since april 2007 [10:03] it's flagged invalid [10:03] yes, I know. even more disturbing IMHO [10:05] just tried it on a lucid test machine. problem still occurs. :( [10:12] that's a bitch [10:12] try reopening the bug [10:16] RoyK: yep, just reopened it/expanded it to cover postgresql-8.4 [10:30] A question about BIND9. Is using wildcard (*) for all subdomains a bad idea? Should I define a line for every subdomain I have? [10:31] elnur: I explicitly define each subdomain [10:31] elnur: though I don't know why wildcard would be bad [10:32] imjustmatthew, then why do you define them explicitly? what is your reasoning? [10:34] I define them each explicitly to make it easier for me to maintain my config files [10:34] some of them are shared accross servers since some are fron-facing DNS and some are LAN facing [10:34] that sentence had to many somes... [10:35] some of the subdomain configs are shared accross servers since some of the servers are front-facing DNS and some are LAN facing [10:35] imjustmatthew, and how does that affect to your choice? I don't get it [10:36] The only problem I know with explicitly defining every subdomain is that you have to wait while global DNS records update for every subdomain you add. [10:39] I've never had a problem with that, but yes, you are right that could be a downside if you change the records frequently [10:41] imjustmatthew, ok. thanks for your help [10:43] putting them in explicitly (and removing them when not used anymore) might be useful to remember what you actually use ;) [11:01] New bug: #580891 in samba (main) "samba filenames wrong" [Undecided,New] https://launchpad.net/bugs/580891 [11:44] how many cpu cores does ubuntu server support? [11:44] as in out of the box [11:53] RoyK, that's likely limited to the 2.6.32 kernel which is likely significantly higher than you'll find available via Intel and AMD processors and related mainboards. [11:53] what is that? 64? 256? [12:01] the out of the box ubuntu 10.04 server kernel looks like CONFIG_NR_CPUS=64 [12:15] should suffice even for a quad 12-core [12:19] should only be a problem for those lucky guys with an 8 socket system [12:37] i am trying to setup a virtual server and am looking for suggestions on what to use. [12:37] xen, kvm or vmware [12:42] resno: Well, unless you have any specific preferenses or reasons, go with KVM. If nothing else because that's what has been closest integrated into Ubuntu. [12:42] penguin42: those opterons are only made for quad usage anyway - not enough hypertransport links for more [12:43] andol: ive heard that it doesnt work on every platform. how can i check to see if itll work? [12:43] im not running a desktop as a server [12:44] im running a desktop as a server [12:45] resno: egrep '(vmx|svm)' --color=always /proc/cpuinfo [12:45] if you get output, your cpu supports kvm [12:45] * resno crosses fingers [12:46] no such luck [12:46] so whats the next best option? [12:46] you're stuck with xen [12:47] resno: Are we talking about always-availible servers, or just something to use for testing and/or development? [12:47] andol: it will be always on. running services for my home network [12:48] Tweeda: and xen is hacky setup from what ive tried on the ubuntu docs [12:48] s/docs/wiki/ [12:48] resno: agreed. I run centos5 for my last remaining Xen host [12:49] Tweeda: is it easier to setup in centos? [12:50] A Xen host? Yes, it's available as an optional kernel and from the same codebase redhat supports commercially. It works. [12:51] Setting up a Xen host (Dom0) in Ubuntu is a bit of work as resno related [12:52] resno: Also note that Debian has a pretty good Xen integration. [12:53] resno: http://wiki.debian.org/Xen [12:53] might also look at a cpu upgrade :) [12:53] * Tweeda hasn't tried debian Dom0 [12:54] Tweeda: maybe, i doubt ill need a better cpu. ill only need more hard drive space :) [12:54] resno, oh, before you take the xen plunge, you might want to check settings in your bios. Often, the factory settings disable virtualization support on the cpu. [12:55] debian has its own version naming conventions, ): [12:55] Tweeda: We run our fair share of that at work. [12:56] andol, then debian would probably be closer to home than centos. [12:56] Tweeda: My thought exactly :) [12:57] resno: But as Tweeda say, first take a peak at your BIOS settings, and see if there's an option to enable hardware virtualization, or so. [12:57] I'd still rather pick up a cpu that supports virt and run kvm [12:57] Tweeda: what cpus support it? [12:58] most new cpus. it's usually listed in the tech specs. well, tech specs worth their salt. [12:58] I run kvm on really cheap amd cpus with little fuss [12:58] what do you consider "new cpus"? [12:59] last few years [12:59] resno: As a very general rule of thumb, if the cpu is 64-bit, it likely also has hardware virtuzliation support. [12:59] RoyK: Oh didn't know the 12 cores were limited to 4 sockets; still you can do 8 socket on Nehalem-Ex and I think thats hyperthread, so you cang et 8x8x2->128 threads [12:59] andol: Except the bottom end ones - like this lapto [12:59] p [13:01] penguin42: I guess 48 cores will suffice for most use :þ [13:02] resno, what CPU are you running? [13:02] any suggestions for which benchmark suite I should use to benchmark a 16-core system? [13:02] I'm curious about memory bandwidth - it's an amd system, so it should be good, but it'd be nice to see the actual numbers [13:02] Tweeda: p4 [13:05] RoyK: It really depends what you want to do with it; I'm always a slave for compile times - make -j is always fun :-) But are you a databasey person or a supercomputery maths person? [13:05] this system is meant for doing modelling [13:05] well then give it one hell of a model [13:05] volcanic ash chewing :þ [13:06] RoyK: You might better fan filters for that :-) [13:06] hehe [13:06] after Eyjafjallajökull started farting, we've had a lot to do at work [13:14] RoyK: what do you work ;-) [13:14] nilu.no [13:17] ah ok :-) [13:18] that new 16-core machine has something like the same cpu power as the other eight boxes they've using for models so far :) [13:28] 64 gigs of ram should probably suffice even for this :þ [15:04] how do I disable samba on lucid? [15:04] doesn't seem to be anything under /etc/rc2.d [15:05] hello has anyone here used a sunfire v100 [15:06] not with ubuntu.... [15:06] but you have used one [15:06] RoyK: /etc/init.d/samba4? [15:06] I think we have a few running s10 [15:07] Tweeda: samba4? [15:07] Tweeda: anyway, the init.d script isn't run directly, mostly from /etc/rc2.d [15:07] ive been given 3 and i need to know how to set them up [15:08] NativeAngels: i have couple of sunfires, why? [15:08] RoyK, I use rcconf, so if /etc/init.d/samba4 exists: rcconf --off samba4 [15:08] NativeAngels: that's sparc, right? [15:08] ive never used rackmounts before [15:09] but always wanted a rackmount [15:09] NativeAngels: the problem isn't that they're rackmounted, but iirc they don't have a graphics card [15:09] and someone was giving them away [15:09] you may have to add a graphics card or install using a serial console [15:10] serial console, of course [15:10] ok [15:10] and how do i do this [15:10] but, if it's a sparc, lucid won't work at all [15:10] sparc is being removed even from ports archives [15:10] it's sparc [15:11] i also go given a sun ultra 5 [15:11] oh [15:11] throw that away [15:11] ATA3 controller and not very fast cpu [15:12] ive been reading a few sites on the v100 it was saying about a console cable [15:13] yeah... if no one steps up to mainatain sparc arch in lucid, it will die by feature freeze [15:13] NativeAngels: yes, you need console cable [15:13] where can i get these from [15:13] hello.. i'm having problems with the kerberos authentication.. i dont know how to access home folders using kerberos and nfsv4 [15:14] would the likes of pc world have one ? [15:14] i've already added the user to the kerberos principals and received a ticket, but can't cd to ~ [15:14] iirc, sun machines have special version :) [15:14] but i might be wrong [15:14] * incorrect goes to his garage and returns with a hand full of cables for NativeAngels [15:16] pc world won't but most geeks will have a garage full of them [15:16] NativeAngels: I guess you can find the pinout somewhere on the web - get a serial cable, clip it, fasten the correct cables to an RJ45 plug [15:17] NativeAngels: or you might want to call Sun^WOracle and buy one from them, it probably won't cost more than $200 or so [15:17] eeks [15:17] Hi there [15:18] DId anyone solved the kjournald2 ext4 problem or is there a known fix for this? [15:18] are you in the uk incorrect [15:18] http://www.sunhelp.org/unix-serial-port-resources/serial-pinouts/ [15:18] NativeAngels, yes [15:18] do you have any of these cables ? [15:18] Am now unpacking 240k email for setting up a mailserver and I can see, that every 5seconds or so it waits unitl some kjournald writeouts [15:19] NativeAngels, i have boxes and boxes of stuff [15:19] me to [15:19] but not them [15:19] best thing to do is subscribe to a local recycle mailing list [15:19] NativeAngels: http://www.sunhelp.org/unix-serial-port-resources/serial-pinouts/ [15:19] or make it as RoyK points out [15:19] more fun to make it [15:20] soldering a D-sub is a bitch, but just using an already soldered serial cable is trivial [15:21] a 9pin is easy [15:21] a 15pin vga is hard(er) [15:21] so no one know ? [15:22] everybody running ext4 without a problem? [15:22] I just wonder, becasue google showed me hundreds of fucked up ppl [15:23] I just installed my first box with ext4 [15:23] works so far, although I haven't been micromonitoring it [15:23] Jeniczek, for a year [15:23] i have 30 servers on ext4 [15:24] its xfs i've had problems with [15:24] well, I am not a LInux guru, but If i clearly see, that unpacking 100gig archive takes more than hour and I cann see taht every 5 seconds kjournald2 and pdflush blocks the unzipping process [15:24] incorrect: out of interest - what problems? [15:24] then I know it is fucked up :( [15:25] http://bbs.archlinux.org/viewtopic.php?id=84953 [15:25] Jeniczek: bad hardware? [15:25] https://bugs.launchpad.net/ubuntu/+bug/442443 [15:25] Launchpad bug 442443 in ubuntu "kjournald2 writes to disk about every 10 seconds" [Undecided,New] [15:25] and so on and so on [15:25] RoyK, just file corruption [15:26] incorrect: that's quite bad [15:26] well, its a 1200Euros PCIE card [15:26] and I havent got any issues on ext3 afaik [15:26] incorrect: I was using xfs a lot earlier, never seen anything like corruption, though [15:26] let me see iotop doesn't show that for me [15:26] and as I can see a lot of ppl around having that problem [15:26] Jeniczek: how much data is on that? [15:26] so its definately on my side [15:27] my ext4 system is fine [15:27] Jeniczek: upgrade from ext3 to ext4? [15:27] RoyK 3 RAIDs, 1 74gig RAID1, 1 2.3TB RAID5 and 1 3TB RAID5 [15:27] no [15:27] clean install [15:27] if you can handle the downtime, downgrade to ext3, that works [15:27] my 15tb array doesn't have problems on ext4 [15:27] but moving 5TB+ will take some time [15:28] well I dunno how to downgrade to ext3 from ext4 workin system [15:28] * RoyK uses zfs for storage [15:28] the RAID is now almost empty [15:28] am now fillin it up [15:28] you can't downgrade, only mkfs [15:28] so if it really is ext4 issue, than now is the rigt time to do it [15:28] hm [15:28] what will ext4 help? [15:28] over ext3? [15:29] ext3 is pretty rock stable [15:29] dunno, ext4 was the default [15:29] so i chose it [15:29] I'd say recreate the filesystems [15:29] move the data [15:29] wait for someone to fix the bugs [15:29] other way round [15:29] move data then recreate fs;) [15:30] can you help me with the recreate command? [15:30] obviously, yes [15:30] am in a biiig hurry, I got one hour :( [15:30] mkfs.ext3 :D [15:30] Jeniczek: move the data to somewhere, mkfs -t ext3 on the now unused fs, move the data back, etc [15:30] well, ill man mkfs then [15:30] i still don't understand what's the problem [15:30] rsync localfile/* remotehost:/somewhere/ [15:30] ivoks: bad i/o performance [15:30] the fact that kernel is flushing data to disk? [15:30] mkfs.ext3 /dev/something or other [15:31] rsync remotehost:/somewhere/ localfile/* [15:31] oh wait i forgot to unmount and mount [15:31] rsync -a [15:31] the problem is, than right now am unzipping the 100GIG archive, and every 5 to 10 seconds a kjournald2 and pdflush stops the whole unzip process, does somethin, and then it keeps goin [15:31] yeah, it's flushing data to the disk :) [15:31] Jeniczek: are any of those filesystems empty? [15:32] nope [15:32] but not full so much [15:32] the mail raid5 is almost empty [15:32] am now setting it up [15:32] Jeniczek: then move the data around to empty one [15:32] then recreate an ext3 fs on that [15:32] then check if you see the same behaviour on ext3 [15:32] /dev/sdc1 1.4T 215G 1.1T 17% /var/mail [15:32] /dev/sdd1 2.7T 48G 2.6T 2% /var/storage [15:32] but it would still need to flush data to the disk [15:32] if you do, it's not an ext4 problem [15:33] otherwise, there wouldn't be anything on the disk :) [15:33] just move the storage over to mail [15:33] ext3 does it every 3 seconds [15:33] well, i need to move the mail to the storage [15:33] ext4 every 5 or 10 [15:33] ivoks so what do you think that causes this problem? [15:33] Jeniczek: there's more than four times as much on mail as there is on storage [15:33] these aren't problems [15:34] that's how systems work [15:34] pile up data in memory [15:34] and then flush it onto disk [15:34] so why is it so slow? [15:34] what sort of drives? [15:34] how many? [15:34] cache on your raid card? [15:34] hardware or software raid? [15:34] hardware [15:34] did you enable cache on it? [15:34] 256MB, BBU included, 3ware 9650 [15:35] 12Ports [15:35] there's tw_cli utility [15:35] is the write cache enabled? [15:35] download it and enable write cache [15:35] how can I found it out? [15:35] tw_cli [15:35] either with tw_cli or rebooting into 3wares bios [15:35] oh god not a 3ware [15:35] I have installed the 3ware utlilty which set ups some kind of web server and allow me to connect to a WEb based gui to control those raids [15:35] what's wrong with 3ware? [15:35] i found my 3ware controller was slower than software raid5 [15:35] I've used 3ware for years - works well [15:35] Jeniczek: you can use that too [15:36] 3ware rulez [15:36] xfs too [15:36] incorrect: software raid usually beats it all [15:36] xfs on 3ware = success :) [15:36] what i have heard, 3ware is a good raid card [15:36] it is [15:36] okey [15:36] not in my experience [15:36] so what should I do now? [15:36] about the cache [15:36] my PERC6's blew the 3wares away [15:36] fuck dell :P [15:36] if you can reboot the box, do so, hit alt+3 during boot [15:37] check if write cache is enabled [15:37] it normally is [15:37] RoyK I can reboot it, but am not in the same city the box is atm, and i dont have a KVM [15:37] so I can get into Alt 3 [15:37] ok [15:37] Jeniczek: you said you installed web utility? [15:37] but I can get to the web based thing [15:37] then you need tw_cli [15:37] anyone has an idea, why nfsv4 maps users to nobody on the client? [15:37] i benchmarked dell, hp, supermicro [15:37] Jeniczek: then use it [15:38] well, ill try to dig fro some cache thing in there now [15:38] out of all the servers dell came top other than for disk, but that was because hp put nearline sas drives in [15:38] JanC: there's controler settings in management [15:39] i put the near line sas drives in the dell and it blew the hp away [15:39] bah [15:39] Jeniczek: ^^ [15:39] Jeniczek: you'll see Unit Write Cache setting [15:40] okey, mmt [15:43] hm [15:43] it seems its enabled :( [15:44] ivoks http://i40.tinypic.com/2d2a9eg.jpg [15:44] is logic disk builded or still building? [15:44] is this it? [15:44] Jeniczek: protection? [15:44] I dunno, thats the system disk [15:44] or the system raid [15:44] you are copying to raid1 or raid5? [15:44]